Comparing the testing effect under blocked and mixed practice: The mnemonic benefits of retrieval practice are not affected by practice format

Mem Cogn (2017) 45:81–92 DOI 10.3758/s13421-016-0641-8 Comparing the testing effect under blocked and mixed practice: The mnemonic benefits of retrieval practice are not affected by practice format Magdalena Abel 1 & Henry L. Roediger III 1 Published online: 27 July 2016 # Psychonomic Society, Inc. 2016 Abstract The act of retrieving information modifies memory in critical ways. In particular, testing-effect studies have demonstrated that retrieval practice (compared to restudy or to no testing) benefits long-term retention and protects from retroactive interference. Although such testing effects have previously been demonstrated in both between- and within-subjects manipulations of retrieval practice, it is less clear whether one or the other testing format is most beneficial on a final test. In two paired-associate learning experiments conducted under typical testing-effect conditions, we manipulated restudy and test trials using either blocked or mixed practice conditions while equating other factors. Retrieval-practice and restudy trials were presented either separately in different blocks (blocked practice) or randomly intermixed (mixed practice). In Experiment 1, recall was assessed after short and long delay intervals; in Experiment 2, the final memory test occurred after a short delay, but with or without an interfering activity before the final test. In both experiments, typical testing effects emerged, and critically, they were found to be unaffected by practice format. These results support the conclusion that testing effects are robust and emerge to equal extents in both blocked and mixed designs. The generality of testing effects further encourages the application of retrieval practice as a memory enhancer in a variety of contexts, including education. Keywords Retrieval practice . Testing effect . Delay . Interference . Practice format * Magdalena Abel 1 Department of Psychology, Washington University in St. Louis, One Brookings Drive, Box 1125, St. Louis, MO 63130-4899, USA The general assumption in education, and often in cognitive psychology, is that the act of testing memory is a neutral affair; tests are given to assess one’s knowledge, not to change it. However, 40 years ago R.A. Bjork (1975) argued that recall not only measures memory but modifies it in several ways, often positively. In recent years, the testing effect—the benefit of the act of recall on later retention—has been frequently studied, and much has been learned. The testing effect refers to the finding that practicing the retrieval of previously studied material can boost memory and enhance long-term retention for the tested materials in comparison to a no-exposure control condition (e.g., Wheeler & Roediger, 1992), or even in comparison to control conditions in which the materials are reread for the same amount of time (e.g., Carrier & Pashler, 1992; Roediger & Karpicke, 2006; for a review, see McDermott, Arnold, & Nelson, 2014). Moreover, retrieval has been shown to protect the practiced material from the detrimental influence of subsequent learning (i.e., from retroactive interference; Halamish & Bjork, 2011; Potts & Shanks, 2012). On the other hand, research on retrieval-induced forgetting indicates that retrieval can also entail negative effects for memory of items related to the retrieved items when retrieval is carried out only selectively (Anderson, Bjork, & Bjork, 1994; see also Anderson, 2003). When retrieval practice occurs for some but not all information, such selective retrieval can cause forgetting of related, but unpracticed contents (relative to a control condition without any practice; for a recent meta-analysis and review on retrieval-induced forgetting, see Murayama, Miyatsu, Buchli, & Storm, 2014). Although the present experiments are concerned with the testing effect, their main motivation was derived from prior work on retrieval-induced forgetting. In addition, the present work also relates to the generation effect, so we briefly review relevant studies from these other domains before providing the rationale for our experiments. 82 The special importance of retrieval as a memory modifier is underscored by studies investigating whether the observed effects of retrieval practice are really specific to retrieval, or might alternatively also arise when memories are strengthened in a different way instead (e.g., by means of restudying). In particular, several studies have documented the point that retrieval-induced forgetting emerges only after retrieval practice, but not after restudy, and of course (by definition) the retrieval-practice effect occurs after testing (and the comparison condition is often a restudy control). Importantly, for both effects, such differences between retrieval practice and restudy have been successfully demonstrated with between-subjects and within-subjects manipulations of the two practice types. For instance, the positive influence of retrieval practice on long-term retention in the testing-effect literature has been demonstrated when different subjects engaged in retrieval practice or restudy (e.g., Karpicke & Roediger, 2008; Pyc & Rawson, 2010), but also when the same subjects practiced some materials by means of retrieval practice and others by means of restudy (e.g., Butler, 2010; Zaromb & Roediger, 2010). Similarly, the negative influence of selective retrieval practice for related but unpracticed material has also been shown to occur when different subjects (e.g., Anderson, Bjork, & Bjork, 2000; Bäuml & Aslan, 2004; Ciranni & Shimamura, 1999) or the same subjects (e.g., Hanslmayr, Staudigl, Aslan, & Bäuml, 2010; Wimber, Rutschmann, Greenlee, & Bäuml, 2009) engaged in selective retrieval practice and/or restudy. Yet another important issue has been less frequently examined: Few studies have investigated whether the format of the practice type matters: Are the effects of retrieval practice affected by whether practice occurs in blocks of trials (e.g., blocks of items are restudied or are tested) or, rather, the two types of trials are mixed together? A recent study by Dobler and Bäuml (2013) was the first to address whether practice format was of relevance for retrieval-induced forgetting. Their results showed that when practice was blocked and retrievalpractice and restudy trials were clearly separated, only selective retrieval practice (but not restudy) led to retrieval-induced forgetting (thus replicating prior work; e.g., Hanslmayr et al., 2010). However, when retrieval-practice and restudy trials were randomly intermixed, retrieval-induced forgetting emerged after both types of practice. Dobler and Bäuml interpreted their results as a reflection of dynamic effects between retrieval practice and restudying when the two types of trials are intermixed. That is, in relating their findings to previous work from the task-switching literature (e.g., Allport, Styles, & Hsieh, 1994; Campbell, 2005; Meuter & Allport, 1999), Do (...truncated)